International Journal of Computer Science Engineering

and Information Technology Research (IJCSEITR)

ISSN(P): 2249-6831; ISSN(E): 2249-7943
Vol. 4, Issue 4, Aug 2014, 127-134
TJPRC Pvt. Ltd.

REALIZATION OF ARITHMETIC OPERATIONS ON SIGNAL SEQUENCES USING

NORMAL ALGORITHMS AND REWRITING CYCLIC NORMAL AUTOMATON
U. M. FERNANDES DIMLO1 & E. G. RAJAN2
1
2

Research Scholar, University of Mysore, Karnataka, India

Founder & President, Pentagram Research Center Pvt. Ltd, Hyderabad, Andhra Pradesh, India

ABSTRACT
Symbolic computation is study and development of algorithms and software for manipulating mathematical
expressions and objects. In signal processing data is represented using symbols. But today computers are basically numeric
computers hence automation of basic and advanced signal processing operations became very difficult task. Even though
effort has been made to write algorithms and programs for signal processing operations but time complexity of such
implementations are very poor. In this paper signals are represented as strings or words which can be stored in computer
memory, further normal algorithms are written to perform basic arithmetic operations like addition, difference, product and
division of two signals using rewriting cyclic normal automata.

KEYWORDS: RCNA, Normal Algorithms, Substitution Formula

INTRODUCTION
Data can be represented either using numbers or symbols. Hence based on data representation there are two types
of computation can be performed on data such as

Numeric computation

Symbolic computation
Computation carried out by human beings before invention of computer is purely symbolic. Once computers

come into existence numeric computation became very popular. Any task which is solved by computers takes data in the
[10] form of numbers. For example looping a certain task for n number of times, doing any bank transaction,
finding integration using Simpsons method are carried out in a numeric way. But there is no significant improvement in
numerical differentiation.
Symbolic computation is study and development of algorithms and software for manipulating mathematical
expressions and objects. In signal processing data is represented using symbols. But today computers are basically numeric
computers hence automation of basic and advanced signal processing operations became very difficult task. Even though
effort has been made to write algorithms and programs for signal processing operations but time complexity of such
implementations are very poor. The problems identified are processing signal processing operations [09] on numeric
computers, difficulty in writing algorithms, Rare availability of software to perform advanced signal processing operations,
Modeling symbolic operations in computer process able form, Poor time complexity of existing methods.

www.tjprc.org

editor@tjprc.org

128

U. M. Fernandes Dimlo & E. G. Rajan

DEFINITION OF IMPORTANT TERMS

Rewriting Cyclic Normal Automata
Let us consider a normal algorithm N consisting of n totally ordered substitution formulas {ui vi: 1 i n} and
its EPT version [9]. We shall define a rewriting cyclic normal automaton for the EPT system in the following manner:
Let Q be the set of states corresponding to the semi-Thue productions [11] contained in the scheme of the EPT
rewriting system where Q = {qij | 0 i n 1; 0 j k: i represents the ith substitution formula, j represents the jth
semi-Thue production of the ith substitution formula [12] and k represents the number of states, that is, the number of
semi-Thue productions corresponding to the ith substitution formula}. For convenience, the state corresponding to the ith
substitution formula is termed as a major state and the states corresponding to the k semi-Thue productions of the ith
substitution formula as minor states. [13]Let ui vi be the ith substitution formula. Then, the number of minor states
corresponding to the ith major state is given by the following rules.

if | ui | = | vi | = lm, then k = 2lm

[NOTE: | ui | is the length of ui]

if | ui | < | vi | then k = 2lm

if | ui | > | vi | then k = 2lm-1

Substitution Formulas
The formula of the form P->Q used in normal algorithms to derive given string is called as substitution formula.
There are two types of substitution formulas

Simple substitution formula

Terminal substitution formula

This formula comprises of generic variables, auxiliary [14] variables and string. Terminal formula will be used to

end the derivation where simple formula used to replace part of a string by another string while derivation.
Normal Algorithms
Finite set of symbols are known as alphabet denoted by A or . A word w is finite sequence of symbols from
. Basic set theory operations and notations like intersection, union, set difference, concatenation, subset, proper subset,
belongs to, not belongs to can be the same in automata also. A variable is said to be generic [15] variable whose values are
taken from alphabet . Usually generic variables are denoted by symbols like, and. In automata null string is
represented by the symbol. A Markov Algorithm over an alphabet A is a finite ordered sequence of productions xy,
where x, y A*. Some productions may be Halt productions. e. g. abc b, ba x (halt).

ADDITION OF TWO SIGNALS

Let s1, s2 be any two signals which represented using strings w1, w2 in automata. Because in automata everything
which has been recognized will be considered as word. Let we convert given signals into integer sequence where for
example integer 3 will be represented using ||| and integer 5 will be represented as |||||. Hence numbers of vertical lines are
equal to given integer sequence of input signal.
Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0

Realization of Arithmetic Operations on Signal Sequences Using Normal Algorithms and Rewriting Cyclic Normal Automaton

Let N

ADD

129

is an RCNA based normal algorithm which is used to add to signals denoted by integer sequence.

The substitution formulas for addition are as follows.

Table 1: Formula for Addition Two Signals
Formula

Formula Number

Comments
Replace symbol + by null string
and stop the addition process

Let us consider two signals which are represented as integer sequence ||| and ||. We have only one formula in
which + is replaced by then if we concatenate any string with null string the result will be the same string. The addition
of these two signals carried out as follows.
Table 2: Addition of Two Signals 3 and 2

The

RCNA

for

the

Step No.
1
2

Transformation
||| + ||
||| ||

|||||

above

EPT

system

is

Formula Used
Input string
Formula 0
Resultant string
(w = w = w)
[1,2]

defined

by

Q= { q00,q10,q20,q21,q30,q40,q50,q51 } with the set of edges T where, T = { (qINI, *, q00),

\/,q20),

CADD

<Q,INI,FIN>

where

(q00, \/, q10), (q10,

(q20,\/,q30), (q20,+/,q21),

(q21,/ ,q00), (q30,\/,q40), (q40,\/,q50), (q50, \/,qFIN), (q50, /,q51), (q51,/ ,q00) }
The graphical representation of CADD is shown in figure 1

Figure 1: RCNA for Addition of Two Signals

DIFFERENCE OF TWO SIGNALS

The signals which are given will be converted into integer sequence p (n) and q (n) where n is length of the
sequence. In general pipeline operator is used to denote signals in automata theory. In finding difference of two signals
three cases may arise which are listed below

Integer sequence of signal s1 is greater than signal s2 sequences

Signal s1 integer sequence is same as s2 sequence

www.tjprc.org

editor@tjprc.org

130

U. M. Fernandes Dimlo & E. G. Rajan

Length of p (n) is shorter than length of q (n)

Let N

SUB

is an RCNA based normal algorithm which is used to add to signals denoted by integer sequence.

The substitution formulas for addition are as follows.

Table 3: Substitution Formula for to Find Difference of Signals

Formula

Formula Number

|-| -

|- |
0

1
2

Comments
Substitute in the place |-| in the given
input string
Replace at last |- with |
If both signals are same result is 0

Let us consider two signals which are represented as integer sequence ||| and ||. We should use formula 0 for twice
then formula 1 or once to get desired result. The subtraction of these two signals carried out as follows.
Table 4: Difference of Signals 3 and 2

The

RCNA

for

the

Step No.
1
2
3

Transformation
||| - ||
||-|
|-

above

EPT

system

is

Formula Used
Input string
Formula 0
Formula 0
Formula 1
Resultant signal
defined

by

CSUB

<Q,INI,FIN>

where

Q={q00,q10,q20,q11,q12,q21,q22,q30,q31,q32,q40,q41,q50,q51,q60,q61,q62,q70,q71,q72} with the set of edges T where,

T = {(qINI, *, q00),( q00, , \/,q10),(q10, , \/,q20),(q10,1/ ,q11),(q11, /,q12),(q12, / ,q00),(q20,-/-,q21),
(q21,/,q22),(q22,/

,q00), (q30,1/,q31),(q31, /,q32),(q32, /

,q00),(q40, \/,q50), (q40, 1/,q41),

(q41, / ,q00),(q50,-/-,q51),(q51, / ,q00),(q60, 1/ ,q61),(q61, /,q62), (q62, / ,q00),(q70,1/1,q71),(q71,-/ ,qFIN),

(q71, /, q72), (q72, / , q00) }
The graphical representation of CSUB is shown in figure 2.

Figure 2: RCNA to Find Difference of Two Signals

Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0

Realization of Arithmetic Operations on Signal Sequences Using Normal Algorithms and Rewriting Cyclic Normal Automaton

131

PRODUCT OF TWO SIGNAL SEQUENCES

Let us consider two non negative integer sequences [3] of any two signals and then convolution of these two is
nothing but multiplication of pair of integers. Let N MUL is an RCNA based normal algorithm which is used to multiply to
signals denoted by integer sequence. We are considering only two positive signals to make the process simple and clear.
Firstly we like convert everything into sequence bs then we can apply successively formula replacing *b by |* which
yields final result. The substitution formula for multiplication is as follows.
Table 5: Substitution Formulas for Product of Two Signals
Formula

Formula Number

a| |ba

b| |b
O*| O*
|*O *O
|*|*a
*| *
a
*b |*
*.

1
2
3
4
5
6
7
8

Comments
Add |ba in the place a| in the
given input string
Write in reverse
Truncate | from substring
Truncate | from substring
Replace | by a
Remove | from string
Empty string at a
Replace by |
. in the place of *

For example let us consider two signals s1, s2 which are denoted by integer sequence x, y where x=||| and y =||.
Let us apply NMUL algorithm to find product of s1 and s2 which is as follows.
Table 6: Product of Signals 3 and 2
Step No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
www.tjprc.org

Transformation
|||||
||a||
|||ba|
|||b|ba
||||bba
|a||bba
||ba|bba
||b|babba
|||bbabba
a||bbabba
|ba|bbabba
|b|babbabba
||bbabbabba
|bbabbabba
bbabbabba
bbbbabba
bbbbbba
bbbbbb
|bbbbb
||bbbb
|||bbb
||||bb
|||||b
||||||
||||||

Formula Used
Input string
4
0
0
1
4
0
0
1
4
0
0
1
5
5
6
6
6
7
7
7
7
7
7
8 (resultant string)
editor@tjprc.org

132

U. M. Fernandes Dimlo & E. G. Rajan

The

RCNA

for

the

above

EPT

system

is

defined

by

CMUL

<Q,INI,FIN>

whereQ={q00,q10,q20,q11,q12,q13,q30,q21,q22,q40,q50,q41,q42,q51,q52,q60,q70,q71,q72,q73,q80,q81,q82,q83,q90,q91,q92,q93,q100,q101,q102,q1
03,q104,q105,q110,q111,q112}

the set of edges T where,

T= { (qINI, *, q00), (q00, , \/,q10), (q10, \/,q20), (q11,|*|/*a,q12),(q12, /,q13),(q13, /

,q00),

(q20, \/,q30), (q20,a|/|ba,q21), (q21, /,q22), (q22, / ,q00), (q30, , \/,q40), (q40, , \/,q50), (q40,a|/|ba,q41),
(q41,/,q42),

(q42,/

(q60,,\/,q70),
(q74,/

(q70,|*|/*a,q71),
,q00),

(q84, /
(q100,,

,q00),

/,q52),(q52,

(q50,b|/|b,q51),(q51,
(q71,a|/|ba,q72),

(q80,|*|/*a,q81),

(q72,a|/|ba,q73),

(q81,a|/|ba,q82),

(q100,*b/|*,q101),

(q101,*b/|*,q102),

(q102,*b/|*,q103),

,q00),

(q73,

/,q74),

(q82,a|/|ba,q83),

,q00), (q90,a|,q91), (q91,a|,q92), (q92,a|,q93), (q93, /,q94), (q94, /

\/,q110),

(q83,/,q84),

,q00), (q90, , \/,q100),

(q103,*b/|*,q104),

(q104,

/,q105),

(q105, / ,q00), (q110,*b/|*,q111), (q111,*b/|*,qFIN), (q111, /,q112), (q112, / ,q00) }

The graphical representation of CMUL is shown in figure 3

Figure 3: RCNA for Product of Signals

DIVISION OF STRING SEQUENCES

Let us consider two signals s1, s2 which are denoted as non negative integer sequence where s1 is factor of s2.
Here the division forced to remainder as zero and only for positive integers. The division of signals can be accomplished
by finding the number of times s2 occur in s1 by factoring method [4,5].
Table 7: Normal Algorithm for Division of Signals
Formula

Formula Number

/ |*|*/

/|*$|

$
|

3
4

Impact Factor (JCC): 6.8785

Comments
Cancel number of |s in dividend which are present in
divisor.
Replace / by empty string
Write | at the end of string after cancelling equal length
substring from divisor and dividend. Where $ is end of
the string.
Replace $ by empty string
Replace | by

Index Copernicus Value (ICV): 3.0

Realization of Arithmetic Operations on Signal Sequences Using Normal Algorithms and Rewriting Cyclic Normal Automaton

133

Let us consider two signals which are represented as integer sequence |||| and ||. We should use formula 0 for twice
then formula 1 or once to get desired result. The division of these two signals carried out as follows.
Table 8: Division of Signals 4 and 2
Step No.
1
2
3
4
5
6
7
8
9
10
11
12
13
The

RCNA

for

the

above

Transformation
||||/||$
||||/||$
||||/||$|
||||/||$|
||||/||$||
|||/||$||
||/||$||
|/||$||
/||$||
||$||
|$||
$||
||
EPT

system

Formula Used
Input string
0
2
0
2
4
4
4
4
1
4
4
3(output string)
is

defined

by

CDIV

<Q,INI,FIN>

whereQ={q00,q10,q20,q30,q40,q31,q32,q50,q41,q42,q43,q44,q45,q60,q51,q52,q53,q54,q55,q60,q61,q62,q70,q71} the set

of edges T where,
T= {(qINI, *, q00), (q00, , \/,q10), (q10, , \/,q20), (q20,

\/,q30),

(q30, \/,q40), (q30,///,q31),(q31, /,q32), (q31, / ,q00), (q40, \/,q50),(q41,///,q42),(q42,||/||,q43),(q43,$/$|,q44),

(q44, /,q45), (q45, / ,q00), (q50, , \/,q60),(q50,||/||,q51),(q51,///,q52), (q52,||/||,q53),(q53,$|/$||,q54),
(q54, /,q55),(q55, /

,q00), (q60, \/,q70),(q60,|/ ,q61), (q61,/| ,q62),(q62,$/ ,qFIN),(q70, /,q71),

(q71, / ,q00) }
The graphical representation of CDIV is shown in figure 4

Figure 4: RCNA for Division of Signals

CONCLUSIONS
In this paper a new technique for performing operations on signals using normal algorithms and RCNA has been
www.tjprc.org

editor@tjprc.org

134

U. M. Fernandes Dimlo & E. G. Rajan

represented, which solves the problems like numeric computation of signals, representation of signals in computer process
able format. The use of automata in signal processing is still in infancy stage. More efforts are required to explore various
methods using automata to implement various signal processing operations. The features of automata will make them as
alternative for designing signal operations as conventional methods. In summary, it is considered that RCNA based automata
have an excellent potential in signal processing.

REFERENCES
1.

Aberth, O, Computable Analysis, McGraw Hill, New York, 1980.

2.

Arbib, M. A, Algebraic Theory of Machines, Languages and Semi groups, Academic Press, New York, 1968.

3.

Barwise, Jon, Handbook of Mathematical Logic, North-Holland Publishing company, 1977.

4.

Beeson, M. J, Foundations of Constructive Mathematics, Springer Verlag, Heidelberg, 1985.

5.

Benthem, J. F.A.K.V, The Logic of Time, D. Reidel, London, 1983.

6.

Bishop, Errett, Foundations of Constructive Analysis, McGraw Hill, 1967.

7.

Bobrow, D.G, Symbol Manipulation Languages and Techniques, Proceedings of the IFIP Working Conference on
Symbol Manipulation Languages, North-Holland Publishing Company, Amsterdam, 1968.

8.

Bridges, D. S, Constructive Functional Analysis, Pitman, London, 1979.

9.

Clifford and Preston, The Algebraic Theory of Semi groups, American Mathematical Society, Providence, 1961.

10. Copi, I. M, Symbolic Logic, MacMillan Publishing Company, 1979.

11. Dalen, D. V, Logic and Structure, Springer Verlag, Heidelberg, 1980.
12. Ershov, Yu. L, and Palyutin, E.A, Mathematical Logic, Mir Publishers, Moscow, 1984.
13. Giardina, C. R, and Dougherty, E.R, Morphological Methods in Image and Signal Processing, Prentice Hall Inc,
1988.
14. Gleason, A. M, Fundamentals of Abstract Analysis, Addison Wesley. 1966.
15. Griswold, R. E, Poage, J. F, and Polonsky, I. P, The SNOBOL4 Programing Language, Prentice Hall, 1971.

Impact Factor (JCC): 6.8785

Index Copernicus Value (ICV): 3.0